The overall aim of the National Heart, Lung and Blood Institute (NHLBI)/Suburban Hospital Cardiovascular MRI Research Project is to develop new approaches in assessing patients with cardiovascular disease with MRI technology. 1) Detection and characterization of acute coronary syndrome with MRI. We found that a rest MRI scan had higher sensitivity and specificity for diagnosing non-ST elevation acute coronary syndrome than cardiac risk factors, ECG, and troponin (Kwong RY et al. Circulation 2003;107:538-544). We extended this work in a second protocol that used adenosine stress MRI to evaluate 141 consecutive patients with troponin-negative acute coronary syndrome. The overall sensitivity and specificity for detecting ischemic heart disease were both greater than 90%. An abnormal adenosine stress MRI had significant 1 year prognostic value. (Ingkanisorn et al. JACC 2006;47: 1427). We determined that gadolinium delayed enhancement cardiovascular magnetic resonance corrrelates with clinical measures of myocardial infarction. Ingkanisorn et al. J Am Coll Cardiol 2004;43:2253-9. This study imaged patients with acute myocardial infarction an average of 2 days post-MI. The transmural extent of delayed enhancement predicted the recovery of regional myocardial function. Determining what myocardium was ischemic during an acute coronary occlusion is potentially valuable clinically and for research purposes. We completed a preclinical project to characterize recently ischemic myocardium and demonstrated that we can image the ischemic area at risk after myocardial perfusion has been restored using manganese contrast agents. (Natanzon A et al. Radiology. 2005 Sep;236(3):859-66. Preliminary experience with a manganese based contrast agent in humans suggested this agent may have too many side effects to be useful in acute imaging. Next, we found that T2-weighted MRI can determine the area at risk in acute MI and is thus complementary to delayed enhancement imaging in acute reperfused MI (Aletras et al. Circulation 2006;113: 1865)and in non-reperfused myocardial infarction (Tilak et al. Investigative Radiology 2008;43: 7-15). Despite the promise of using T2-weighted images in diagnosing the ischemic area at risk associated with acute coronary syndrome, commercial software used to obtain these images is limited by artifacts and other technical factors that reduce the diagnostic accuracy to about 70%, an unacceptable level of performance. We developed 2 new methods for imaging myocardial edema associated with acute coronary syndrome (Kellman et al Magn Resonance Med 2007;57: 891) and a second method that may offer better signal to noise ratio but is more difficult to implement (Aletras AH et al Magn Reson Med 2008;59: 229). The T2-prep SSFP method markedly improved the diagnostic accuracy of determining the coronary territory associated with an acute myocardial infarction (Kellman et al Magn Resonance Med 2007;57: 891). The T2-prep SSFP method also works well in humans and correlates with coronary angiographic measures of area at risk (Berry C et al. Circ Cardiovasc Imaging. 2010;3:527-535). 2) Characterizing myocardial infarction and viability with MRI. We developed a phase sensitive reconstruction method which improves the quality of heart attack images and minimizes the influence of user selected parameters on the apparent size of the heart attack (Kellman P et al. Magnetic Resonance in Medicine 2002;47:372-383 and Kellman et al. Magn Reson Med 2004;51:408-12). Our histopathological validation of the phase sensitive reconstruction method and a validation study showing that a computer algorithm can accurately measure infarct size (Hsu L et al. J Magn Reson Imaging. 2006;23(3):309-14) and the method works in humans (Hsu L et al. J Magn Reson Imaging 2006;23:315-22). We extended this work to characterize IL-2 myocarditis and community acquired myocarditis. We developeed a method to improve the contrast between myocardial infarction and the blood in the left ventricular cavity (Kellman et al. J Magn Resonance Imaging 2005;22: 605-613) and a method that allows imaging myocardial infarction without resorting to breatholds that are not feasible in all patients (Kellman et al. 2005;53: 194-200). Non-rigid motion corrected methods have additional benefits over the first generation registration programsa that we developed (Ledesma-Carbayo MJ et al. J Magn Reson Imaging 2007;57:891-897). Most recently, we have validated that gadolinium based contrast agents correlate with infarct-related fibrosis at resolutions approaching a cellular level (Schelbert E et al. in press). We have applied these advanced imaging methods in clinical and epidemiological studies of myocardial infarction. We found strong relationships between infarct size and left ventricular remodeling (Hirsch GA et al. Am J Geriatr Cardiol 2007;16: 222-228). We recently completed analysis of the ICELAND MI study of the prevalence of myocardial infarction in 978 participants in the AGES-Reykjavik study. Preliminary results indicate a much higher than expected prevalence for unrecognized myocardial infarction than suggested by prior epidemiological studies relying on ECG methods. 3) We have been working on improving first pass myocardial perfusion imaging through careful quantitative analysis. We have shown that the MRI can measure myocardial perfusion as accurately as microsphere injections (a gold standard method only usable in animal models) Christian TF, Rettmann D, Aletras AH, Liao S, Taylor JL, Balaban RS, Arai AE. Absolute Myocardial Perfusion by MRI Using a Dual-Bolus First-Pass Method: Benefits Over Qualitative and Semi-Quantitative Analysis. Radiology 2004;232:677-84. We confirmed that most of the conclusions derived from Dr. Christian's work applies to humans (Hsu et al. JMRI 2006;23: 315). We completed a study in humnas comparing the diagnostic accuracy of quantitative stress perfusion with quantiative coronary angiography. This work, which is in review, indicates that the fully objective quantitative stress perfusion methods can diagnose patients with significant coronary stenosis without needing additional input from delayed enhancement images or cine MRI. We also developed methods for accelerating perfusion image acquisition (Kellman et al. Mag Resonance Med 2004;52: 200-204) and used a variant of these methods to measure T2* during the first pass of contrast through the heart (Kellman et al. Mag Resonace Med 2006;56: 1132-4). This later paper is important since T2* is thought to be a major source of image artifacts and false positive findings on MRI perfusion studies. In fact, T2* does not appear affected severely enough to explain the endocardial artifacts seen on many patients. Correction of non-linearity between signal intensity and gadolinium concentration improves the accuracy of perfusion quantification (Hsu L et al. J Magn Reson Imaging 2008;27: 793-801). We have also recently studied the reproducibility of quantitative perfusion analysis (Christian et al. J Magn Reson Imaging 2008;27: 1271-1277). 4) Characterization of myocardial abnormalities. Beyond infarction and ischemia, many disease processes alter the characteristics of myocardium. We have developed methods for separating water and fat in cardiac MR images to improve the diagnosis of arrhythmogenic right ventricular dysplasia (Kellman P et al. Magn Reson Med. 2009;61:215-21).